1
1
Граф коммитов

3722 Коммитов

Автор SHA1 Сообщение Дата
Nathan Hjelm
40845d36bc Merge pull request #1045 from hjelmn/openib_fixes
btl/openib: fix argument type for internal atomic function
2015-10-20 14:32:59 -06:00
Nathan Hjelm
55d24ee7a3 btl/openib: fix argument type for internal atomic function
This was fixed on my btl 3.0 branch but the changeset got lost in a
rebase. Fixes issues with lock ups when using osc/rdma.

References open-mpi/ompi#1010

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-10-20 13:47:28 -06:00
Nathan Hjelm
6d3041335f opal/keyval: reset buffer pointer/size in finalize
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-10-20 13:10:44 -06:00
Nathan Hjelm
9602484568 Merge pull request #1040 from hjelmn/mtl_priority
Change how cm's priority is calculated
2015-10-19 14:18:36 -06:00
Nathan Hjelm
8b5810f7f7 mca/base: add priority output to mca_base_select
The mca_base_select function uses returned priorities to select the
best component/module. This priority may be of use to the caller so
pass that information back in an optional argument. If the priority is
not needed pass NULL.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-10-19 12:32:41 -06:00
Ralph Castain
363f62a506 Fix singleton operations when running under a SLURM allocation. Sadly, SLURM's PMI will return success even if the PMI server isn't actually available. This leads to erroneous selection of pmix and ess components. So add a further requirement (namely, that we see a job_step envar) to the SLURM pmix components along with some modification of ess selection code to avoid the problem 2015-10-17 20:24:03 -07:00
Jeff Squyres
a23ab2cfba Merge pull request #1031 from jsquyres/pr/hwloc-1.11.1
hwloc: upgrade from v1.11.0 to v1.11.1
2015-10-15 12:52:37 -04:00
Jeff Squyres
f69364e768 hwloc: upgrade from v1.11.0 to v1.11.1
Taken from upstream v1.11.1 release.

Fixes open-mpi/ompi#981.
2015-10-15 08:58:33 -07:00
Jeff Squyres
aceb1ebb47 Merge pull request #1026 from hjelmn/static_mutex
opal static mutex initializers
2015-10-14 22:10:51 -04:00
Jeff Squyres
a7e7ecd42d opal_progress_thread: fix stale comment 2015-10-14 18:25:31 -07:00
Nathan Hjelm
039c7dbcd6 opal/mutex: add static mutex initializers
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-10-14 16:08:41 -06:00
Howard Pritchard
3521a0f9ff Merge pull request #1016 from hppritcha/topic/fix_issue_1005
btl/ugni: fix very poor aries bw problem
2015-10-14 10:10:28 -06:00
Howard Pritchard
eaba98ce5d btl/ugni: fix very poor aries bw problem
The handling of RDMA get alignment in ugni BTL for Aries
(cray xc) was wrong, resulting in very poor bandwidth
for ugni BTL on aries.

Verified using osu_bw now gives sensible bandwidth on
Aries.

Fixes #1005

Signed-off-by: Howard Pritchard <howardp@lanl.gov>
2015-10-13 16:01:17 -05:00
Jeff Squyres
71dfba9ed3 Merge pull request #845 from ggouaillardet/topic/pmpi_vs_mpi
Remove --enable-mpi-profile configure option (i.e., always build PMPI bindings)
2015-10-13 12:13:10 -04:00
Jeff Squyres
12e796dcaf hwloc: headers are not in $includedir
They are in $opalincludedir.  Use the neutral "$pkgincludedir", which
will get translated under the covers to opalincludedir.
2015-10-13 05:59:52 -07:00
Gilles Gouaillardet
291a464efb configury: remove the --enable-mpi-profiling option
and directly call the PMPI_* symbols from C and Fortran bindings
2015-10-13 08:52:35 +09:00
Jeff Squyres
1dfb685f8d libevent: headers are not in $includedir
They are in $opalincludedir.  Use the neutral "$pkgincludedir", which
will get translated under the covers to opalincludedir.
2015-10-12 15:20:59 -07:00
Jeff Squyres
e8b4a4c800 wrappers: add support for opallibdir, opalincludedir 2015-10-12 15:18:31 -07:00
rhc54
db467d1c8f Merge pull request #969 from annu13/master
sync with PMIX master
2015-10-09 16:04:20 -07:00
annu13
cc5e1e26a5 sync with pmix master (repo_rev git69c398e) 2015-10-09 15:17:43 -07:00
Nathan Hjelm
90db00e37f Merge pull request #996 from hjelmn/openib_progress_thread
btl/openib: remove extra threads
2015-10-08 07:31:27 -06:00
Nathan Hjelm
b8af310efa btl/openib: remove extra threads
This commit removes the service and async event threads from the
openib btl. Both threads are replaced by opal progress thread
support. The run_in_main function is now supported by allocating an
event and adding it to the sync event base. This ensures that the
requested function is called as part of opal_progress.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-10-07 12:30:41 -06:00
Rolf vandeVaart
952d01db70 Add volatile to workaround GCC optimization bug in 4.9.2 and greater. Thanks to oere for fix 2015-10-07 12:56:58 -04:00
annu13
5787e9248f cleaned up debug stmts 2015-10-06 06:25:36 -07:00
annu13
30ba00e05d sync with master 2015-10-06 06:04:54 -07:00
annu13
4c371ea6d7 Merge remote-tracking branch 'upstream/master' 2015-10-06 06:03:37 -07:00
Nathan Hjelm
59aa93e1b6 opal/mpool: add support for passing access flags to register
This commit adds a access_flags argument to the mpool registration
function. This flag indicates what kind of access is being requested:
local write, remote read, remote write, and remote atomic. The values
of the registration access flags in the btl are tied to the new flags
in the mpool. All mpools have been updated to include the new argument
but only the grdma and udreg mpools have been updated to make use of
the access flags. In both mpools existing registrations are checked
for sufficient access before being returned. If a registration does
not contain sufficient access it is marked as invalid and a new
registration is generated.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-10-05 13:53:55 -06:00
Rolf vandeVaart
81346949b8 Merge pull request #975 from rolfv/pr/fix-6.0-compile-error
Fix typo that caused compile error
2015-10-05 13:11:35 -04:00
Nathan Hjelm
e0d9e6553f mpool: mpi_show_mpi_alloc_mem_leaks don't print out message if there are no leaks
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-10-05 10:48:56 -06:00
Rolf vandeVaart
fc83400123 Fix typo that caused compile error 2015-10-05 11:44:40 -04:00
annu13
6f37c0e3e8 sync with PMIX master 2015-10-02 17:25:48 -07:00
annu13
7434c47626 sync with PMIX master 2015-10-02 17:17:48 -07:00
Gilles Gouaillardet
dc883cff8d opal/util: fix parse_ipv4_dots prototype 2015-10-01 14:03:08 +09:00
Ralph Castain
8f6855459d Cleanup some coverity warnings 2015-09-30 10:33:53 -07:00
Nathan Hjelm
3c33a8e94b btl/ugni: adjust exclusivity below sm and vader
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-09-29 15:40:35 -06:00
Ralph Castain
ec5d001538 Don't set the return value pointer to NULL as it actually is required to point to real storage - just return an error code if a modex recv doesn't succeed. 2015-09-28 20:45:50 -07:00
Gilles Gouaillardet
2ac09d5a8d pci: do not probe PCI topology on Solaris unless effective uid is root
Otherwise libpciaccess sends a big error message to stderr:
  Error opening /devices/pci@0,0:reg: Permission denied

(cherry picked from commit open-mpi/hwloc@d93c7c0960)
2015-09-29 09:42:58 +09:00
Nathan Hjelm
12bd300c40 Merge pull request #929 from hjelmn/add_procs
Update add_procs support
2015-09-28 17:29:13 -06:00
Ralph Castain
a4a3dfd480 Cleanup the code a bit by simply adding our nspace to the top of the list of jobid <-> nspace correlations. Add two new APIs to opal_pmix for registering new jobid/nspace pairs and retrieving an nspace given a jobid - these are required to support connect/accept. No impact on the PMIx library. 2015-09-28 08:50:13 -07:00
Ralph Castain
f713e71d51 Minor cleanup - add jobid <-> nspace in one more place 2015-09-27 14:48:39 -07:00
Ralph Castain
fad5638596 Resolve the naming issue when direct-launched by PMIx-enabled RMs using a minimal-impact approach. Detect if we were launched via ORTE - if so, then use our standard methods for computing the jobid. If not, then just hash the nspace to create the jobid, and track the jobid <-> nspace correspondece down in the opal/mca/pmix/pmix1xx component. We then do the translation any time a function that passes process names is invoked. 2015-09-27 09:57:59 -07:00
Todd Kordenbrock
3e63a3458c portals4: add support for dynamic add_procs() to all Portals4 components
In the default mode of operation, the Portals4 components support
dynamic add_procs().

The Portals4 components have two alternate modes (flow control and
logical-to-physical) that require knowledge of all procs at startup.
In these modes, mtl-portals4 sets the MCA_MTL_BASE_FLAG_REQUIRE_WORLD
flag and btl-portals4 sets the MCA_BTL_FLAGS_SINGLE_ADD_PROCS flag
to tell the PML that we need all the procs in one add_procs() call.
2015-09-24 22:12:57 -05:00
Todd Kordenbrock
3afac9e37d btl-portals4: fix PMIx integration problem
After PMIx integration, the thrid parameter to OPAL_MODEX_RECV() is
opal_process_name_t instead of opal_proc_t.  This commit replaces
proc with &proc->proc_name.
2015-09-24 21:53:20 -05:00
Ralph Castain
0140ff048d Now that we have an "isolated" PLM component, we cannot just let rsh silently decline to run when it cannot find a launch agent - if we do, then we will -always- run on the local node. So if the user specifies a launch agent and we can't find it, then generate a pretty error message, report a fatal error back to the component select, and exit out.
This required modifying the mca_component_select function to actually check the return code on a component query - it was blissfully ignoring it.

Also do a little cleanup to avoid bombarding the user with multiple error messages.

Thanks to Patrick Begou for reporting the problem
2015-09-24 07:16:48 -07:00
Ralph Castain
209600fe26 Sync to PMIx master 2015-09-23 21:00:30 -07:00
Nathan Hjelm
a31ea2e780 Merge pull request #932 from hjelmn/ompi_coverity
btl/openib: fix udcm coverity errors
2015-09-23 19:06:22 -06:00
Nathan Hjelm
60f3dbd160 btl/openib: fix udcm coverity errors
Fix CID 1312120: Uninitialized scalar variable

The response type will always be set unless a message of another type is passed to this function. To make sure that error is caught I am adding an assert.

Fix CID 1312116: Dereference after null check

This is a potential bug. If there is no endpoint data for an incoming connection a rejection should be sent. In this case we would just SEGV.

Fix CID 1312115: Dereference after null check

Clear error in the error message. Use the queue pair number that was passed in.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-09-23 17:08:27 -06:00
Nathan Hjelm
54a4061d88 Add support for detecting when dynamic add_procs is not possible
This commit adds support to the pml, mtl, and btl frameworks for
components to indicate at runtime that they do not support the new
dynamic add_procs behavior. At the high end the lack of dynamic
add_procs support is signalled by the pml using the new pml_flags
member to the pml module structure. If the
MCA_PML_BASE_FLAG_REQUIRE_WORLD flag is set MPI_Init will generate the
ompi_proc_t array passed to add_proc from ompi_proc_world () instead
of ompi_proc_get_allocated ().

Both cm and ob1 have been updated to detect if the underlying mtl and
btl components support dynamic add_procs.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-09-23 16:22:05 -06:00
Ralph Castain
749bd4e6fe Plug a few memory leaks identified by valgrind 2015-09-23 15:21:04 -07:00
Nathan Hjelm
5cb44c2827 Merge pull request #926 from hjelmn/pmix_fix
pmix: do not touch recently freed memory
2015-09-23 09:55:36 -06:00